Oops reading scanlations now means I’m in the mental rabbit hole of thinking where bottlenecks of cultural transfer occur, and how fan translation was a niche world decades ago. Now having better information infrastructure and cross referencing translations immediately means that the bottleneck still exists, but is exponentially larger. I also appreciate that the fan translation sphere can obviously recognize quality and has a visceral extreme distaste for shit machine translation.
The idea that it is impossible to translate poetry is not an indictment on the quality of translation, but is a testament to the insufficiency of language as a whole. Language has limits in what it is able to communicate, and poetry is an attempt to explore those limits. The intent of the poet is crucial to the piece being translated. Getting the “correct” translation would require ascertaining the “correct” intent, and the whole endeavor falls apart with such an unreasonable expectation. Humans fail at it all the time, and even picking translation notes is an artistic decision in itself. I have no doubt that machine translation will get to the point of quality far beyond the worst fan translation, but I don’t think that it’s too far a stretch to say that the shitty fan translation still has value. That someone would apply a skill they tenuously have a handle on, to create something that did not exist before. To stake your claim, “this piece is important to me, and I want to share it with others”. Yes, machine translation offers an opportunity for curatorial democracy, but we lose the piece of the individual, the flawed person who spent hours considering what the authors’ intentions were, and released them for others to read. A great loss.